Clinical Testing on External Ventricular Drainage Guidance using Mixed Reality
Abstract Augmented reality surgery systems have played an important role in assisting physicians in their operations. However, applying the system to brain neurosurgery is challenging. In addition to using the augmented reality technology to display the 3D position of the surgical target position in real time, we also need to consider the display of the scalpel entry point and scalpel orientation, and their accurate superposition on patients. This paper proposes a mixed reality surgical navigation system, which accurately superimposes the surgical target position, scalpel entry point and scalpel direction on a patient's head and displays it on a tablet, facilitating the visual and intuitive way for the brain neurosurgery. Based on the current neurosurgery navigation system, we integrated mixed reality technology on it. We first independently tested the accuracy of the optical measurement system-NDI Polaris Vicra, and then designed functions that a physician can quickly point out the surgical target position and decide an entry point position, and a tablet can display the superimposed images of surgical target, entry point, and scalpel, and perform the correctness of scalpel orientation. Then we used the Dicom of the patient CT to create a phantom and it’s AR model, imported this AR model into the APP, and installed and executed the APP on the tablet. In the preoperative phase, the technician first superimposed the AR image of head and the scalpel in 5-7 minutes, and then the physician point out and set the target position and entry point position in 2 minutes on a tablet, which then dynamically displayed the superimposed image of the head, target position, entry point position, and scalpel (including the scalpel tip and scalpel spatial direction). We successfully conducted multiple experiments on phantom and six experiments on clinical neurosurgical EVD practice. In the 2D-plane-superposition model (n = 60), the optical measurement system (NDI Polaris Vicra) was feasible of the visualization space with high accuracy (mean error ± standard deviation (SD): 2.013 ± 1.118 mm). In the clinical trials in the hospital (n = 4), the average technician preparation time was 6.317 minutes. The average time (n = 4) required for the physician to set the target position and the entry-point position and accurately overlay the orientation with a surgical stick was 3.5 minutes. In the preparation phase, the average time required for the Dicom image processing and program importing was 120 ± 30 minutes. The designed mixed reality optical surgical navigation system can successfully achieve clinical accuracy, and guide physicians to perform brain surgery visually and intuitively. In addition, the physician can use the APP of the tablet device to instantly obtain Dicom images with the designated patient, change the position of the surgical entry point, and instantly obtain the accurate surgical path and surgical angle after the modification. This design can be used as the basis for various AR or MR brain surgery navigation systems in the future.